date: 22:16:17, October 15, 2018
my ip: (
your ip: ()
CCBot/2.0 (
Welcome, my name is Emmanuel Roberto-Richiardone, this is my place on the web.

You can contact me at
e AT

My PGP/GPG public key
More about me here
main blog
software projects
hardware concepts
texts docs tips
freebsd tips
linux tips

outside links:
My photograph gallery

Friends pages:
  NetStudent association
  tollari retrocomputing
  Jovi Berton

Web-things often followed:
  Supersonic Electronic
  Juxtapoz Magazine
  Wooster Collective
  Internal Bleeding
  Cyanide and Happiness
  oddity central
  the oat meal
  QDB @

RSS feed enabled!

  October 2018  
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31        

search with google
all the web

. o .
. . o

multi page
Samba 3 share with authentication with a Windows Active Directory Service on Debian 6Tue, 22 Jan 13
Foreword: hosts in question are:

All following commands are done as root on Debian. Starts installing required packages.
apt-get install samba krb5-user libpam-krb5 winbind
EDIT /etc/hosts  windowsdomain
EDIT /etc/krb5.conf
    default_realm = NETWORK.LOCAL
    kdc_timesync = 1
    ccache_type = 4
    forwardable = true
    proxiable = true

        kdc =
        admin_server =
        default_domain = NETWORK.LOCAL

Verify that you can logon correctly with user Administrator on the Active Directory:
# kinit Administrator@NETWORK.LOCAL
# klist
Note: Clock skew limits are configurable in the Kerberos protocols. The default setting is five minutes.
EDIT /etc/samba/smb.conf
        workgroup = NETWORK
        realm = NETWORK.LOCAL
        security = ADS
        syslog = 0
        log file = /var/log/samba/log.%m
        max log size = 1000
        idmap uid = 10000-20000
        idmap gid = 10000-20000
        winbind enum users = yes
        winbind enum groups = yes
        template homedir = /home/%D/%U
        template shell = /bin/bash
        client use spnego = yes
        client ntlmv2 auth = yes
        encrypt passwords = yes
        winbind use default domain = yes
        restrict anonymous = 2
        valid users = @"Domain Users"
        winbind refresh tickets = yes

        comment = file server
        path = /share
        read only = No
        hosts allow =
        case sensitive = No
        preserve case = Yes
        available = yes
        browseable = yes
        create mask = 0664
        directory mask = 0775
Restart daemons:
/etc/init.d/winbind stop
/etc/init.d/samba restart
/etc/init.d/winbind start
Verify that you can join the Active Directory, using Administrator and password "adminpass"
net ads join -U Administrator%adminpass
EDIT /etc/nsswitch.conf
passwd: compat winbind
group:  compat winbind
shadow: compat winbind

hosts:     files dns wins
networks:  files dns

protocols:   db files
services:    db files
ethers:      db files
rpc:         db files

netgroup:     nis
Restart daemons:
/etc/init.d/winbind stop
/etc/init.d/samba restart
/etc/init.d/winbind start
Verify that since here things works with:
net ads info
wbinfo -u
getent passwd
getent group
ADD in /etc/pam.d/common-auth
auth sufficient
ADD in /etc/pam.d/common-session
session required
You can check the login of a user of the domain with:
# login
Finally some cosmetics, set permissions of the share:
chmod 775 /share
chgrp "Domain Users" /share
Mount a single partition from a complete disk imageFri, 28 Sep 12
Suppose you've done an image dump of a complete disk (i.e. /dev/sdb) with a command like this:
root@n2:~# dd if=/dev/sdb of=/tmp/mydump.dd bs=16k
Now you want to mount a partition included in the disk image. First know how was made the disk with fdisk:
root@n2:~# fdisk -ul /tmp/mydump.dd
You must set cylinders.
You can do this from the extra functions menu.

Disk /tmp/mydump.dd: 0 MB, 0 bytes
64 heads, 63 sectors/track, 0 cylinders, total 0 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000

                    Device Boot      Start         End      Blocks   Id  System
/tmp/mydump.dd1   *          63     3330431     1665184+   6  FAT16
OK, you see there's one partition and it starts at cilinder 63. So calculate at which byte of the image disk the partition begins:
root@n2:~# echo 63*512 | bc
And finally mount the partition from the image dump, skipping the bytes to ignore:
root@n2:~# mount -o ro,loop,offset=32256 -t vfat /tmp/mydump.dd /mnt
Debian Lenny on NSLU2Tue, 11 Jan 11
Sone issue with this guide, please do not follow it! Required material:
  • a Linksys NSLU2 device (I had a 10/2006 266MHz version)
  • a GNU/Linux computer
  • a USB drive/pendrive with minimum 1 GB of size
  • some network cables and a switch
  • access to internet
  • 2 hours of time

Credits for this guide goes to cedric of There is also a famous automatic procedure reported here, but for me it doesn't works as after the installation it reboots and always restarts the debian-installer; also I prefer knowing what I am doing.
Firstly install the utility upslug2 on a GNU/Linux computer to remotely set up the NSLU2:
$ sudo apt-get install upslug2
Download a flash image, it's a minimal Debian that will be used as a starting place for the final installation. You can take this one from the guide mentioned above.
Start the NSLU2 in the flash mode. To do so, power down the NSLU2 and remove USB drivers. Then, you have to keep the reset button while powering on the NSLU2; after something like 10 seconds, the Status LED change color from green to red, and you have to release the reset button. The status LED starts blinking. (note: on my NSLU2, the LED was not clearly visible, so I powered on the board whitout the casing).
Connect together through the switch the NSLU2 and the GNU/Linux computer. Now launch as root on the GNU/Linux computer:
$ sudo upslug2
You should see the MAC address of the NSLU2 to flash. If not, repeat the reset procedure.
Load the minimal Debian image (caution to not detach power and network cables):
$ sudo upslug2 -i debianslug-3.10-beta.bin
LED blinks while flashing. When done, you should see an output that ends with "Rebooting... done"; the NSLU2 reboots and the minimal Debian boots. It is ready when you hear a "beep" tone and the LED is solid green.
The gain access to the system use SSH. The IP address is the one you set up on the Linksys firmware, or the default The user and password also are the one you previously set, or are the default "root" and "opeNSLUg".
$ ssh root@
On the NSLU2, set up with turnup new parameters or confirm the old ones. I set a new static IP address
# turnup init
After restart with those new parameters the NSLU2:
# reboot
When the device is ready (LED stops blinkin), login again to the NSLU2:
$ ssh root@
Plug in your USB drive in the port you want where you will install Debian. Begin to set up the drive, choosing at least of 256MB of swap for your system. Set the first partition as bootable. Take care on mounted partition as the installer has auto-mount feature enabled.
# fdisk /dev/sda
It appears something like:
Disk /dev/sda: 2048 MB
   Device Boot    Start       End    Blocks   Id  System
/dev/sda1   *           1         ??      ??  83  Linux
/dev/sda2              ??        250      ??  82  Linux swap
# mkfs.ext3 /dev/sda1
# mkswap /dev/sda2
Set the NSLU2 telling him with turnup that you want to boot your USB partition:
# umount /dev/sda1
# turnup disk -i /dev/sda1 -t ext3
Set up the installation and start it (it'll takes some times and download files from internet):
# mount /dev/sda1 /mnt/
# swapon /dev/sda2
# mkdir /mnt/work
# cd /mnt/work
# wget
# ar -x debootstrap-udeb_1.0.26_all.udeb
# tar zxvf data.tar.gz
# export DEBOOTSTRAP_DIR=`pwd`/usr/share/debootstrap
# ./usr/sbin/debootstrap --arch arm lenny /mnt/
Mount proc, copy some files and configure some vital information:
# mount -t proc proc /mnt/proc
# cp -a /lib/modules/* /mnt/lib/modules
# cp /etc/resolv.conf /mnt/etc/resolv.conf
# cp /etc/hosts /mnt/etc/hosts
# cp /etc/hostname /mnt/etc/hostname
# cp /etc/network/interfaces /mnt/etc/network/interfaces
# echo "ixp400_eth dev_max_count=1" >> /mnt/etc/modules
Set up the disk information:
# vi /mnt/etc/fstab
proc		/proc		proc	defaults	0  0
/dev/sda1 	/ 		ext3 	defaults 	1  1
/dev/sda2 	swap	 	swap 	defaults 	0  0
Set up a headless system
# vi /mnt/etc/inittab
commenting the following lines:
 #1:2345:respawn:/sbin/getty 38400 tty1
 #2:23:respawn:/sbin/getty 38400 tty2
 #3:23:respawn:/sbin/getty 38400 tty3
 #4:23:respawn:/sbin/getty 38400 tty4
 #5:23:respawn:/sbin/getty 38400 tty5
 #6:23:respawn:/sbin/getty 38400 tty6
and adding at the end the line:
 S0:2345:respawn:/sbin/getty 115200 ttyS0
Chroot to target system and continue the setup:
# chroot /mnt
# passwd root Update packages informations and set the time:
# apt-get update
# apt-get -y install ntpdate
# ntpdate
# tzconfig
Set automatically the time:
# echo "ntpdate -s" > /etc/cron.daily/ntpdate
# chmod 755 /etc/cron.daily/ntpdate
Install some fundamentals:
# apt-get -y install nslu2-utils tune2fs openssh-server
Disable the check of the filesystem at boot:
# tune2fs -c 0 -i 0 /dev/sda1
Verify that in the SSH configuration you can login as root:
# vi /etc/ssh/sshd_config
Look at the line:
 PermitRootLogin yes
Exit and reboot:
# exit
# reboot
If all is done right it should restart on the USB drive and available with SSH when the LED stops blinking:
@ ssh root@
Continue installing with apt-get the services you need as a normal Debian Lenny. Enjoy!
startx at boot without login managerWed, 29 Jul 09
I wanted a computer with Debian without keyboard to automatically load a graphical interface. The solution I was looking for had to be the lightest possible, without frills nor using a login manager like GDM and SLIM.
You cannot start Xorg (or XFree86) from a script automatically loaded at boot; that's because you need a complete shell ambient to execute startx or xinit.
A simple solution could be to start a shell over a tty terminal without asking for the login, an then start in the profile file of the shell the graphical command. Be careful of this configuration, because it starts automatically a (possible buggy) shell at boot

In Debian, in the file /etc/inittab comment out the line
1:2345:respawn:/sbin/getty 38400 tty1
with the following line, replacing USER as necessary:
1:2345:respawn:/bin/login -f USER tty < /dev/tty1 > /dev/tty1 2>&1
Now at boot the computer automatically logon the user USER. Usually in Debian you are using bash. So add the command to start Xorg, startx near the end of the file
A final note: to execute a graphical program in your automatically started Xorg, add the line exec COMMAND in the file /home/USER/.xinitrc.
Ubuntu + Thunderbird + LightningSat, 10 Jan 09
Siccome siamo nel 2009, era ora aggiornassi la mia postazione Ubuntu 7.10 alla 8.10 (e pensare che FreeBSD l'aggiorno almeno una volta al mese). Questa volta e` andata discretamente bene.
L'unico problema riscontrato e` stato con Thunderbird caricando l'estensione per il calendario Lightning, ma anche Sunbird. In effetti dopo l'update, il mio Thunderbird con Thunderbird 0.9 non caricava piu` la schermata dei calendari, ne` era possibile aggiungere nuovi eventi o caricare calendari CalDAV remoti.
La soluzione sta nell'installare una libreria mancante poiche` erroneamente considerata obsoleta:
Per l'installazione e` possibile utilizzare synaptic, oppure il comando
# sudo apt-get install libstdc++5
Nei meandri dei NAS: recupero estremo di RAID5Mon, 6 Oct 08
Premessa: in azienda teniamo i dati cartografici su uno dei tanti NAS in commercio, che rendono accessibili da SMB/CIFS, NFS, FTP il filesystem, su cui gira chissa` quale GNU/Linux nascosto all'utente e configurabile da una scarna interfaccia web.

Sopra avevamo 3 dischi da 500GB SATA-2 (la marca e modello sono volutamente omessi...) in RAID5, per un totale di 930GB spazio disponibile, in teoria sicuri. Si` perche` sempre in teoria, se un disco si rompe, lo si sostituisce con un altro e siamo salvi.

A neanche un anno dall'acquisto dei dischi, il numero 2 si rompe (la definizione migliore e` "fa` sklang-sklang"). Poco male, ne ordino uno nuovo, tempo 3 giorni arriva.

Se non fosse che il giorno dopo il numero 3 viene escluso dal RAID, senza un motivo apparente, dall'interfaccia web non c'e` alcuna spiegazione.

Tanto per capire cosa c'e` che non va`, collego il numero 3 su un PC con sopra System Rescue CD, e scopro subito dal boot che il disco ha circa 120 settori danneggiati. Bene.

Aspetto il disco nuovo, che non tarda ad arrivare. Provo a ricostruire il RAID sul NAS, niente da fare, sempre con zero spiegazioni sul perche`. Provo a sostituire l'elettronica del disco che "fa` sklang" con quella di un altro, il risultato non cambia.

A questo punto mi ritrovo con un RAID5 di 3 unita` di cui una e` persa, e l'altra ha settori danneggiati.

Sul PC di prima, che ormai era li` aperto sul tavolo da due giorni, collego il disco numero 1 buono, il disco numero 3 con settori danneggiati ed il disco nuovo. Verifico che il RAID e` stato fatto dal NAS con mdadm. Provo a ricostruire il RAID a mano ignorando i vari avvisi. Peccato che mdadm ternima la ricostruzione quando incontra il primo settore danneggiato, segnalandomi giosamente che il RAID e` perso per sempre poiche` incontra errori di I/O.

Maledizione pero` anche se qualche file e` corrotto (sulla fine sono solo 120 settori su 1048576000!), voglio recuperare gran parte dei dati!!
Allora mi viene in mente la cosa piu` pazza di questa settimana: fare una copia con errori del disco 3 con settori danneggiati su un altro disco nuovo, e ricostruire il raid con questo, il disco 1 salvo e il disco nuovo di prima (con un totale di 5 dischi uguali di cui 2 rotti).

Ecco la procedura:
A) Collego al PC il disco 3 con settori danneggiati e un disco nuovo, e avvio system rescue cd
B) Faccio una copia con ddrescue scrivendo al posto dei settori danneggiati di origine degli zeri sul disco destinazione:
# dd_rescue -A /dev/sda /dev/sdb
(nota: il processo per i miei 500GB con i settori danneggiati ci ha messo circa 6 ore) Ottengo un disco immagine del 3 con errori
C) Spengo il sistema. Collego il disco 1 che si era salvato, il disco immagine 3 con errori, e l'altro nuovo disco vuoto. Avvio system rescue cd.
D) Ricostruisco il RAID5 ora che non si puo` piu` lamentare degli errori di I/O. Esamino i dischi con fdisk e scopro che il RAID e` sulla seconda partizione dei dischi:
# echo p | fdisk /dev/sda
infatti leggo il tipo "linux raid autodetect" sulla partizione 2. Recupero la stringa ARRAY che mi serve nel file di configurazione:
# mdadm -v --examine --scan >> /etc/mdadm.conf
controllo il file di configurazione /etc/mdadm.conf e verifico che sia sensato, aggiungo ai device anche il nuovo disco vuoto /dev/sdc ottenendo una riga del tipo:
ARRAY /dev/md0 level=raid5 num-devices=3 UUID=XXXX-.....-XXXX
lancio la ricostruzione:
# mdadm --assemble --run --force --update=resync /dev/md0 /dev/sda2 /dev/sdb2 /dev/sdc
e infine verifico lo stato di avanzamento del processo:
# cat /proc/mdstat
E) Quando ha concluso (in /proc/mdstat leggo "[UUU]"), scopro che sopra il RAID non c'e` un filesystem, e quindi ipotizzo ci siano dei volumi logici con LVM. Faccio uno scan del volumi fisici presenti:
# pvscan
trovo in effetti un volume di nome "vg0" (e` indicato dopo "VG "), e provo ad attivarlo:
# vgchange --ay vg0
F) Funziona, ora mi trovo in /dev/vg0/ diversi volumi logici. Trovo quello che piu` mi interessa, il piu` grande, analizzando la dimensione con fdisk:
# ls /dev/vg0/
# echo p | fdisk /dev/vg0/lv1
E` lui, tipo ext3, faccio un check del filesystem:
# fsck -t ext3 -y /dev/vg0/lv1
Ci mette alcune ore (e` pur sempre un'unita` da quasi 1TB), correggendo molti campi sballati degli inode.
G) Monto il filesystem:
# mount -t ext3 -r /dev/vg0/lv1 /mnt/custom
Ok, ho recuperato i 300GB di dati su /mnt/custom, magari qualche file e` corrotto, ma e` *moooolto* meglio di niente. I dati sono stati copiati da rete su un nuovo NAS di qualita` forse superiore, con 3 dischi da 1TB in mirroring (RAID1) con spare (cambiando marca chiaramente).

Morale della storia: Qualsiasi RAID e` inutile, senza spare. (Any RAID, without spare, is useless)
Fedora over UbuntuSun, 27 Jan 08
Want to install Fedora 7 or 8 in a virtual machine on Ubuntu Gutsy?
You have tried VMWare and cannot install it? Fedora freeze in loading and/or cannot find virtual disk?

Well, this is my today problem!
It seems that those version of Fedora don't like too much the buslogic virtual SCSI controller used by qemu and vmware. So, whenever you are using an IDE virtual disk, it still find the virtual controller.

To solve this problem in VMWare, edit the .vmx file of your virtual machine and add
scsi0.virtualdev = "lsilogic"
to use the lsilogic virtual controller, or if you wish to not use SCSI at all add
scsi0:0.present = "FALSE"
to disable any virtual SCSI controller.

top   multi page
apache php i prefer firefox W3C html 4.0 compliant W3C css compliant
This website do not store non-essential cookies.
Emmanuel Roberto Richiardone (e AT richiardone DOT eu)

page viewed 2138 times and generated in 0.001595 s